Digital Futures, in partnership with The University of Manchester, The Christabel Pankhurst Institute, ECH Alliance and Health Innovation Manchester, hosted AI in Healthcare: Transforming the Future with Large Language Models. The event, hosted at Manchester-based Citylabs 1.0 on 24th July 2025, brought together NHS and NHS-related staff, SMEs, academics and researchers, innovators, and those passionate about improving health and care. It offered attendees a unique opportunity to connect and collaborate, and discuss the impact of AI in the healthcare industry.
Ian Hall, Professor of Mathematical Epidemiology and Statistics at The University of Manchester, was present as the Chair of the event. Ian introduced Sarah Barton Townsend, Strategy and Operations Manager of Digital Futures, who delivered a short introduction on its five research themes. Sarah gave prominence to the Digital Health theme, which focuses on using digital technologies to enhance patient care, streamline healthcare processes, and promote preventive health measures. She also gave attendees the chance to look back at previous Greater Manchester Health Ecosystem events in the Digital Health theme.
Ben Latimer, Partnerships Manager, and Caroline Allen, Senior Partnerships Lead, of Accurx gave the first presentation of the day, Integrating Ambient Voice Technology into Clinical Pathways.
This presentation highlighted how NHS staff are spending nearly a third of their day on administrative tasks, hindering the time that they can spend delivering care to their patients. Ambient Voice Technology allows clinicians to record and transcribe interactions with patients, streamlining the process and giving staff this time back to support their patients. Accurx Scribe has over 2,000 NHS providers using its medical device and 30,000 scribes are being completed each week. Ben and Caroline ran a live demonstration, showing attendees how Scribe can benefit clinicians in managing their time efficiently.
Goran Nenadic, a Professor in the Department of Computer Science, of The University of Manchester, gave the second presentation of the day, titled Validation of Clinical LLMs: Opportunities, Risks and Good Practice.
Goran discussed the challenges of validating clinical LLM's, with hundreds of models being available to clinicians. The risk of language model errors in suggesting differential diagnoses for emergency departments was also highlighted, with research suggesting that this is on par with clinicians. However, with 87% of results being accurate, it leaves 23% of errors which underline the question of if these errors are being made in life-threatening or rare cases.
Yutong Zhang, Innovation Adoption Project Manager (iTap), of Manchester University NHS Foundation Trust (MFT), presented next, with the topic of The Innovative Technology Adoption Programme (iTap).
Yutong's presentation covered how iTap is a structured innovation pipeline, helping MFT adopt innovative technologies faster and more effectively. The core of iTap is to prioritise innovations to address the most important challenges within MFT, such as streamlined patient flow to reduce hospital admissions and a reduction of patient waiting time, as well as increased productivity, such as freeing up staff time and improving the care that MFT delivers. Yutong also informed attendees of the benefits of partnering with iTap, with it providing a Trust-wide framework to streamline NHS adoption of innovations that are market-ready.
Joshua Harris, Applied AI Researcher in Advanced Analytics, UK Health Security Agency, provided the fourth presentation of the day, titled UKHSA – Deploying and Evaluating LLMs to Support Public Health.
Joshua's presentation evaluated LLMs in two ways; for public health classification and extraction tasks and its ability to answer questions about UKHSA guidance, looking at potential risks and performance-related issues. Discrepancies included Text Extraction and Preprocessing, Accuracy, Relevant Text Retrieval, Recommendation Classification, and Discrepancy Classification. Joshua also spoke about an interesting publication, that focused on if LLM's can detect gastrointestinal issues in online restaurant reviews. This study tested the abilities of LLM's to look through large volumes of reviews and identify potentially abnormal outbreaks of gastrointestinal numbers.
Anusha Venkatesan and Lorraine Dube, both AI Airlock Regulatory Specialists in Innovative Devices, of the Medicines and Healthcare products Regulatory Agency, provided the final presentation of the day, titled Balancing Innovation and Oversight: AI Airlock and Its Role in Shaping AI as a Medical Device regulation.
This presentation covered the AI Airlock Programme, which explored the regulatory challenges of AIaMD by working directly with real-world products and prototypes. Lorraine covered three variations; Simulation Airlock, Research/ Virtual Airlock, and Real world Airlock. Anusha discussed the AI Airlock workflow programme, which narrowed submissions down from 40 to four pilot cohort candidates across many regulatory challenges. This was followed by a testing phase to understand each product in depth, an implementation stage to test products in real-world and virtual environments, and a review. Results are now being analysed from this programme.
Ben Brown, Co-Founder and Chief Medical Officer of PATCHS, could not attend in-person on the day, but sent over a presentation on the topic of AI to Support Triage in Primary Care, covering how PATCHS is using AI in triage to perform various actions to support patients and clinicians. You can watch this below via YouTube.
The event wrapped up with a breakout session, where Ian asked groups their take on the challenges of developing and deploying AI in the NHS, and a networking lunch quickly followed. This allowed attendees to informally discuss the day's presentations and speak to the incredible speakers. We hope that attendees enjoyed the event and were able to learn more about how AI is beneficial to clinicians and the NHS as an organisation, but also consider the risks that AI poses with possible discrepancies and inaccurate readings.
We want to thank all the speakers, attendees, partners, and staff who made this event possible. We look forward to continuing these important discussions and collaborations in future GM Connected Health Ecosystem events.
Follow the Digital Futures Eventbrite to never miss a GM Connected Health Ecosystem event!
Watch the event recording here
Get In Touch
The University of Manchester's wider Digital Futures network is highly interdisciplinary and operates across the whole range of the University’s digital research - connect with us and keep in touch:
Digital Futures: digitalfutures@manchester.ac.uk | LinkedIn | Website | Sign up to Newsletter | YouTube | BlueSky| Eventbrite